**Dysfunction of Salivary Glands, Disturbances in Salivary Antioxidants and Increased Oxidative Damage in Saliva of Overweight and Obese Adolescents**

**Anna Zalewska 1,\*, Agnieszka Kossakowska 1, Katarzyna Taranta-Janusz 2, Sara Zi ˛eba 3, Katarzyna Fejfer 4, Małgorzata Salamonowicz 4, Paula Kostecka-Socho ´n 4, Anna Wasilewska <sup>2</sup> and Mateusz Maciejczyk 5,\***


Received: 17 January 2020; Accepted: 14 February 2020; Published: 17 February 2020

**Abstract:** Obesity is inseparably connected with oxidative stress. This process may disturb the functioning of the oral cavity, although the effect of oxidative stress on salivary gland function and changes in the qualitative composition of saliva are still unknown. Our study is the first to evaluate salivary redox homeostasis in 40 overweight and obese adolescents and in the ageand gender-matched control group. We demonstrated strengthening of the antioxidant barrier (↑superoxide dismutase, ↑catalase, ↑peroxidase, ↑uric acid, ↑total antioxidant capacity (TAC)) with a simultaneous decrease in reduced glutathione concentration in saliva (non-stimulated/stimulated) in overweight and obese teenagers compared to the controls. The concentration of the products of oxidative damage to proteins (advanced glycation end products), lipids (malondialdehyde, 4-hydroxynonenal) and DNA (8-hydroxydeoxyguanosine) as well as total oxidative status were significantly higher in both non-stimulated and stimulated saliva as well as plasma of overweight and obese adolescents. Importantly, we observed more severe salivary and plasma redox alterations in obese adolescents compared to overweight individuals. In the study group, we also noted a drop in stimulated salivary secretion and a decrease in total protein content. Interestingly, dysfunction of parotid glands in overweight and obese teenagers intensified with the increase of BMI. We also showed that the measurement of salivary catalase and TAC could be used to assess the central antioxidant status of overweight and obese adolescents.

**Keywords:** oxidative stress; antioxidants; saliva; salivary biomarkers; obesity

#### **1. Introduction**

Overweight/obesity is a social problem worldwide, characterized by an increase in body weight that results in excessive accumulation of fat [1]. In recent years, we have observed a steady growth in the frequency of overweight and obesity, observed not only among adults, but also children and adolescents [2]. This results from various genetic, environmental and economic (easy access to cheap and highly calorific food) factors as well as evolutionary conditioning (sedentary lifestyle, low physical activity, low energy expenditure) [1,3]. According to the latest WHO report, nearly 41 million children under 5 years of age are overweight or obese [4]. Interestingly, studies have shown that about 40% of overweight children will continue to gain weight during the puberty period, and about 80% of these obese teenagers will remain obese as adults [5]. Although obesity rates are higher in developed countries, more overweight or obese children live in developing countries, and this trend also applies to European countries.

It has been shown that obesity is associated with an increase in oxidative stress (OS). OS is a condition characterized by disturbed balance between the amount of reactive oxygen species (ROS) produced by the body and activity/concentration of antioxidants responsible for neutralizing ROS [6]. ROS are chemically reactive molecules which, when unbalanced, lead to oxidative modifications of proteins, lipids, carbohydrates and nucleic acids, resulting in obesity-related complications.

It should be emphasized that obesity has been recognized as a major underlying factor in the pathogenesis of serious OS-related health problems, such as hyperlipidemia [7], insulin resistance [8], hypertension [9], type 2 diabetes [10], cardiovascular diseases [7] and certain types of cancer [11]. Overweight/obesity has been shown to adversely affect the condition of the oral cavity, including salivary gland function [12–14]. The pathogenesis of salivary gland lesions in the course of obesity is not fully understood, although the influence of OS is emphasized. It has been demonstrated that adult morbid obesity is associated with disorders of antioxidant systems [15,16] and oxidative damage to salivary proteins, lipids, and DNA [17], while bariatric treatment generally lowers the levels of salivary oxidative damage. However, it does not rescue antioxidant capacity of non-stimulated and stimulated saliva [16,17].

Recently, more and more attention has been paid to the use of saliva in laboratory medical diagnostics, particularly in connection with pediatric diseases, as the collection of saliva samples is non-invasive and thus acceptable to children. It has been shown that proteins and other substances are transported to saliva from blood via the passive process of diffusion, ultrafiltration and active transport. The concentrations of numerous substances in saliva can be correlated with their concentrations in blood plasma, allowing for the use of saliva as an alternative diagnostic material. Moreover, the use of oxidative stress biomarkers is proposed in diagnosing patients with obesity due to the observed changes in enzymatic and non-enzymatic antioxidant levels as well as accumulation of protein and lipid oxidation products in plasma and saliva of obese patients [16,17].

There have been numerous studies on OS in saliva and blood of overweight and obese adults [15–18], whereas no research has been conducted to evaluate salivary redox markers and their usefulness in diagnosing adolescents with excessive body weight. Therefore, the aim of our work is to evaluate antioxidant systems as well as oxidative modifications of proteins and lipids in non-stimulated and stimulated saliva of overweight and obese adolescents.

#### **2. Materials and Methods**

#### *2.1. Patients*

The study was approved by the Ethics Committee at the Medical University of Bialystok, Poland (permission number R-I-002/43/2018). After explaining the purpose and methodology of the study to patients and their parents, written informed content was obtained from each parent/legal guardian.

Our study included adolescents aged 11–18: overweight with BMI z-score ≤ + 1 + 2 < SD (*n* = 20, 10 teenagers) and obese with BMI z-score ≥ + 2 SD (*n* = 20, 10 teenagers).

The control group consisted of adolescents (*n* = 40, 20 teenagers) with normal body weight (BMI z-score < –1 + 1 <, matched by age and gender to the study group.

The adolescents included in the study group were recruited in the Department of Pediatrics and Nephrology of the Medical University of Bialystok, during routine follow-up visits, after performing a dental and biochemical blood test and meeting the conditions for inclusion/exclusion to the study.

The adolescents included in the control group were recruited during dental follow-up visits in the Children's Outpatient Dentistry Clinic (Specialist Dental Centre of the Medical University of Bialystok), initially based on BMI index and a health survey. Then, after obtaining the written consent of participants and their legal guardians, biochemical blood tests were performed. If the participants met the conditions for inclusion/exclusion in the study, they were finally included in the control group.

Patients from the study and control group were qualified by the same experienced pediatrician (K.T.J.) as well as a pediatric dentist (A.Z.). The control group consisted of patients of the Children's Outpatient Dentistry Clinic because, in the Pediatrics and Nephrology Clinic, there were only a few healthy controls that met the inclusion criteria for the study.

Clinical data of the subjects are presented in Table 1.


**Table 1.** Clinical characteristics of patients and healthy controls.

Body weight, height and head and chest circumferences were measured by standard methods. BMI was calculated as weight (kg) divided by the square of height (m2). BMI z-scores that reflect the standard deviation score (SD) for age- and gender-appropriate BMI distribution, were calculated according to the LMS method [19], using reference values from the WHO study [20]. Based on the international norms from the World Health Organization for age- (with an accuracy of 1 month) and gender-specific BMI, BMI cut-offs for children over 5 years of age were the following: overweight–BMI z-score ≤ + 1 + 2 < SD; obesity–BMI z-score ≥ + 2 SD [21].

The inclusion criteria were: adolescents of both sexes (in the case of girls, those who had had menarche) with full permanent dentition. On the day of material collection, the teenage girls were in the first phase of their menstrual cycle.

Patients with deciduous and mixed dentition, with gingivitis (gingival index, GI > 0.5) and pathological changes in the oral cavity mucosa were excluded from the study. Negative general medical history was a necessary factor to qualify for the experiment. The questionnaire completed by the patients included: infectious, autoimmune and metabolic diseases (type 2 diabetes) as well as hypertension, insulin resistance and diseases of respiratory, cardiovascular, digestive, genitourinary and coagulation systems. The exclusion criteria also covered inappropriate behavior and/or refusal to cooperate with the examiner. At least 3 months before the study, patients and the healthy controls had not taken any oral non-steroidal anti-inflammatory drugs, glucocorticosteroids, vitamins, other supplements, or antibiotics. The participants were non-smokers and did not drink alcohol more frequently than once a month.

#### *2.2. Blood Collection*

Blood was collected on an empty stomach, during routine examinations in the case of adolescents from the control group, and for the study groups: during admission to the Pediatric and Nephrology Clinic of the Medical University of Bialystok. Blood was collected in the amount of 10 mL using an S-Monovette® EDTA K3 tube (Sarstedt, Nümbrecht, Germany). After collection, the blood was centrifuged (3000 g, 10 min, 4 ◦C). No hemolysis was observed in any of the obtained plasma samples. Blood cell mass was rinsed 3 times with 0.9% NaCl, and then underwent osmotic lysis using 50 mM cold phosphate buffer (pH 7.4) 1:9 (v/v) [17]. To prevent sample oxidation, 0.5 M BHT (Sigma-Aldrich, Saint Louis, MO, USA; 10 μl/ml blood) was added to the plasma and red blood cell lysate [17]. Plasma and blood lysate were frozen (–82 ◦C). The samples were stored deep-frozen for no longer than 6 months.

#### *2.3. Saliva Collection*

Non-stimulated and stimulated saliva was collected by the spitting method between 7 and 8 a.m., one day after admission to the Pediatrics and Nephrology Clinic of the UMB or during a routine dental check-up. The time since the last meal, tooth-brushing and taking medications was at least 10 h. Samples were collected in a separate room to ensure comfort for the subjects. After rinsing the mouth with water, participants spat out non-stimulated saliva accumulated at the bottom of the oral cavity for 15 min. Stimulated saliva was collected after a 5-min break. Stimulation was performed by dropping 20 μl of 2% citric acid on the tongue every 20 s for 5 min. Both types of saliva were collected in test tubes placed on ice. To prevent sample oxidation, 0.5 M BHT (Sigma-Aldrich, Saint Louis, MO, USA; 10 μl/ml blood) was added to the saliva. After collecting the samples, the volume of saliva was measured in a calibrated pipette with accuracy of 100 μL and saliva flow rate was estimated. Saliva samples were centrifuged (3000 g, 20 min, 4 ◦C, MPW 351, MPW Med. Instruments, Warsaw, Poland) and then frozen (–82 ◦C). Frozen samples were stored for no longer than 6 months.

In each of the obtained saliva samples, the concentration of transferrin was determined by the ELISA test to identify samples contaminated with blood. Transferrin was not detected in any of the saliva samples (data not shown).

#### *2.4. Dental Examination*

Immediately after saliva collection, dental examination was performed by the selected dentists (K. F., M. S., P. K.-S.) in artificial light, using a mirror, an explorer and a periodontal probe (WHO, 621) in accordance with the WHO criteria. The dental examination included the evaluation of DMFT (decayed, missing, filled teeth) and GI (gingival index). The interrater reliability for DMFT was *r* = 0.92, and for GI was *r* = 0.94.

#### *2.5. Biochemical Determination*

The performed assays included: antioxidant enzymes (salivary peroxidase (Px), EC 1.11.1.7, catalase (CAT), EC 1.11.1.6 and superoxide dismutase (SOD), EC 1.15.1.1), non-enzymatic antioxidants (reduced glutathione (GSH), uric acid (UA) and albumin), redox status (total oxidant status (TOS), total antioxidant capacity (TAC) and oxidative stress index (OSI)), advanced glycation end products (AGE), malondialdehyde (MDA), 4-NHE-protein adduct (4-HNE) and 8-hydroxy-D-guanosine (8-OHdG). All results were standardized to mg of total protein. The total protein content was determined using the

bicinchoninic acid (BCA) method and bovine serum albumin (BSA) as a standard (Thermo Scientific PIERCE BCA Protein Assay (Rockford, IL, USA).

In the saliva samples, we analyzed all redox biomarkers. In erythrocytes, antioxidant enzymes were assayed, while in the blood plasma we evaluated non-enzymatic antioxidants, redox status and oxidative damage products as well as interleukin-6 (IL-6) concentration. All assays were performed in duplicate samples (TOS in triplicate samples), and the absorbance/fluorescence of the samples was measured with an Infinite M200 PRO Multimode Microplate Reader (Tecan).

#### *2.6. Antioxidant Enzymes*

The activity of salivary peroxidase (Px, E.C. 1.11.1.7) was determined colorimetrically according to Mansson-Rahemtulla et al. [22] based on the reduction of 5,5'-dithiobis-(2-nitrobenzoic acid) (DTNB) to thionitrobenzoic acid, which then reacted with thiocyanate anions (SCN- ) formed as a result of potassium thiocyanate (KSCN) oxidation by Px. The absorbance was measured at a 412 nm wavelength. The activity of catalase (CAT, E.C. 1.11.1.6) was determined by the colorimetric method described by Aebi [23], based on the measurement of the hydrogen peroxide (H2O2) decomposition rate in phosphate buffer at pH 7.0. The absorbance was measured at 240 nm wavelength. One unit of CAT activity was defined as the amount of the enzyme that decomposes 1 mM H2O2 for 1 min. The activity of superoxide dismutase-1 (SOD, E.C. 1.15.1.1) was determined colorimetrically according to Misra and Fridovich [24] based on the measurement of cytoplasmic activity of the SOD subunit in the reaction of inhibiting the oxidation of epinephrine to adrenochrome at a 320 nm wavelength. It was assumed that one unit of SOD activity inhibits epinephrine oxidation in 50%.

#### *2.7. Non-Enzymatic Antioxidants*

The concentration of reduced glutathione (GSH) was determined colorimetrically based on DTNB reduction to 2-nitro-5-mercaptobenzoic acid under the influence of GSH contained in the assayed samples [25]. Absorbance changes were measured at 412 nm wavelength. Uric acid (UA) concentration was determined by the colorimetric method using a set of ready-made reagents (QuantiChrom TM Uric Acid Assay Kit DIUA-250, BioAssay System Harward, CA, USA). The method is based on the reaction of 2,4,6-tripyridyl-s-triazine with iron (3+) ions in the presence of UA contained in the samples. Changes in the absorbance of the resulting complex were measured at a 590 nm wavelength.

#### *2.8. Redox Status*

Total antioxidant capacity (TAC) was determined colorimetrically as described by Erel [26], based on the measurement of the ability to neutralize the radical cation ABTS ·<sup>+</sup> [2,2-azino-bis-(3-ethylbenzothiazoline-6-sulfonate)] under the influence of antioxidants contained in the tested samples. Changes in the absorbance of ABTS ·<sup>+</sup> solution were measured at a 660 nm wavelength. Total oxidant status (TOS) was determined using the colorimetric method described by Erel [27], based on the oxidation of iron (2+) ions to iron (3+) ions in the presence of oxidants contained in the sample, followed by the detection of Fe3<sup>+</sup> ions by xylene orange. TOS concentration was calculated from the standard curve for hydrogen peroxide and presented in nM H2O2 equivalent/mg total protein. TOS determination was performed in triplicate samples. The oxidative stress index (OSI) was presented as the quotient of TOS to TAC and expressed in % [28,29].

#### *2.9. Products of Oxidative Damage to Proteins and Lipids*

The content of protein advanced glycation end products (AGE) was determined fluorimetrically by the method described by Kalousová et al. [30] based on the measurement of fluorescence of furyl-furanyl-imidazole (FFI), carboxymethyl lysine (CML), pyraline and pentosidine at the excitation wavelength 350 nm and emission wavelength 440 nm. To determine the AGE content, the samples were diluted in PBS buffer (0.02 M, pH 7.0) in a volume ratio of 1:5, and mixed thoroughly [31]. AGE content was determined in duplicate samples and expressed in fluorescence arbitrary units AFU/mg

total protein. The concentration of malondialdehyde (MDA) was determined colorimetrically using thiobarbituric acid (TBA) [32]. The MDA reaction with TBA produces a colored adduct with the maximum absorption at 535 nm wavelength. The concentrations of 4-HNE and 8-OHdG were assessed by ELISA with commercial sets (Cell Biolabs, Inc. San Diego, CA, USA; USCN Life Science, Wuhan, China, respectively), following the manufacturer's instructions included in the package.

#### *2.10. Statistical Analysis*

Statistical analysis of the results was performed using GraphPad Prism 8 and Microsoft Excel 16.16.12 for MacOS. The D'Agostino–Pearson test and Shapiro–Wilk test were used to assess the distribution of the results. Individual groups were compared using the analysis of variance (ANOVA) followed by Tukey's honest significant difference test (Tukey's HSD test). The multiplicity adjusted p value was also calculated. Correlations between redox biomarkers were assessed based on the Pearson correlation coefficient. The results were presented as mean and standard deviation (SD) using tables or graphs. Diagnostic usefulness of the redox biomarkers was evaluated by means of receiver operating characteristic (ROC) analysis. Statistical significance was assumed at *p* ≤ 0.05.

The number of subjects was determined based on our previous experiment, assuming that the power of the test would equal 0.9.

#### **3. Results**

#### *3.1. General Characteristics*

Stimulated secretion was significantly lower in the group of overweight and obese adolescents compared to the controls (↓40%, *p* < 0.0001; ↓51%, *p* < 0.0001, respectively). Teenagers with obesity secreted considerably less saliva after stimulation than their overweight peers (18%, *p* = 0.004). The secretion of non-stimulated saliva did not differ significantly between the study groups as well as in comparison with the control group (Table 2).


**Table 2.** Salivary flow rate, total protein and stomatological findings.

NWS- non-stimulated salivary flow rate, SWS- stimulated salivary flow rate, TP- total protein, C- control, OWToverweight, OB- obese, DMFT= decay, missing, filling teeth, PBI- papilla bleeding index, GI- gingival index, \* *p* < 0.05 vs. C; # *p* < 0.05 vs. OWT.

Total protein concentration in stimulated saliva of overweight and obese adolescents was significantly lower than in the control group (↓33%, *p* = 0.0001; 40%, *p* < 0.0001, respectively). Protein concentration in non-stimulated saliva did not differ significantly between the study groups and compared to the control group (Table 2).

Scatter plots for BMI and NWS/SWS flow rate are presented in Figure 1.

**Figure 1.** Scatter plots for BMI and non-stimulated and stimulated salivary flow rate in healthy children (**A**) as well as overweight and obese adolescents (**B**). BMI- body mass index; NWS- non-stimulated whole saliva; SWS- stimulated whole saliva.

There were no significant differences in the dental indexes DMFT and GI between the controls and the groups of overweight and obese adolescents (Table 2).

#### *3.2. Enzymatic Antioxidants*

The activity of SOD in non-stimulated and stimulated saliva of overweight adolescents was significantly higher than in the control group of adolescents with normal weight (↑60%, *p* < 0.001; ↑48%, *p* = 0.002, respectively). Similar significant differences were observed in the group of obese adolescents, in whom SOD activity in non-stimulated and stimulated saliva was significantly higher than in the control group (↑125%, *p* < 0.001; ↑78%, *p* < 0.001, respectively). There were no differences in salivary SOD activity between overweight and obese subjects. SOD activity in erythrocytes of both overweight (↓59%, *p* < 0.001) and obese (↓58%, *p* < 0.001) adolescents was considerably lower than in erythrocytes of teenagers with normal body weight, and did not differ between the study groups.

The activity of CAT in non-stimulated saliva of obese adolescents was significantly higher compared to the controls (↑75%, *p* < 0.001) and overweight adolescents (↑75%, *p* < 0.001).

The activity of CAT in stimulated saliva of overweight (↑62%, *p* < 0.001) and obese (↑90%, *p* < 0.001) adolescents was significantly higher than in the control group. CAT activity in erythrocytes of obese teenagers was considerably lower than in the control group (↓49%, *p* < 0.001) and in overweight adolescents (↓38%, *p* = 0.02).

The activity of Px in non-stimulated saliva did not differ between the study groups and the controls. Px activity in stimulated saliva and erythrocytes of overweight (↑78%, *p* < 0.001; ↑153%, *p* < 0.001, respectively) and obese (↑57%, *p* < 0.001; ↑153%, *p* < 0.001, respectively) adolescents was significantly higher than in saliva and erythrocytes of the control group (Figure 2).

*J. Clin. Med.* **2020**, *9*, 548

**Figure 2.** Enzymatic antioxidants in overweight and obese adolescents as well as healthy controls. C- control, CAT- catalase, NWS- non-stimulated whole saliva, OB- obese, OWT- overweight, Pxsalivary peroxidase, SOD- superoxide dismutase, SWS- stimulated whole saliva. Differences statistically significant at: \* *p* < 0.05, \*\* *p* < 0.005, \*\*\* *p* < 0.0005.

#### *3.3. Non-Enzymatic Antioxidants*

The concentration of GSH in non-stimulated and stimulated saliva of overweight (↓47%, *p* < 0.001; ↓26%, *p* = 0.005, respectively) and obese (↓65%, *p* < 0.001; ↓54%, *p* < 0.001, respectively) adolescents was significantly lower than in the control group. However, the GSH concentration in overweight adolescents was significantly higher only in stimulated saliva compared to obese subjects (↑38%, *p* < 0.001). Plasma GSH concentration in obese adolescents was significantly lower than in the control group (↓29%, *p* < 0.006).

The concentration of UA in non-stimulated saliva of obese adolescents was significantly higher than in the controls (↑37%, *p* < 0.001). UA concentration in stimulated saliva of overweight and obese adolescents was considerably higher than in the control group (↑157%, *p* < 0.001; ↑198%, *p* < 0.001, respectively). Plasma UA concentration in overweight and obese adolescents was significantly elevated compared to their peers with normal body weight (↑43%, *p* < 0.001; ↑45%, *p* < 0.001, respectively) (Figure 3).

**Figure 3.** Non-enzymatic antioxidants in overweight and obese adolescents as well as healthy controls. C- control, GSH- reduced glutathione, NWS- non-stimulated whole saliva, OB- obese, OWToverweight, SWS- stimulated whole saliva, UA- uric acid. Differences statistically significant at: \*\* *p* < 0.005, \*\*\* *p* < 0.0005.

#### *3.4. Redox Status*

In overweight and obese adolescents, TAC in non-stimulated (↑110%, *p* < 0.001; ↑122%, *p* < 0.001, respectively) and stimulated (↑62%, *p* < 0.001; ↑56%, *p* < 0.001, respectively) saliva as well as plasma (↑61%, *p* < 0.001; ↑75%, *p* < 0.001, respectively) was considerably higher than in the control group.

TOS in non-stimulated (↑113%, *p* < 0.001; ↑288%, *p* < 0.001, respectively) and stimulated (↑115%, *p* < 0.001; ↑170%, *p* < 0.001, respectively) saliva as well as plasma (↑103%, *p* < 0.001; ↑97%, *p* ≤ 0.001, respectively) of overweight and obese adolescents was significantly raised compared to the control group. TOS in non-stimulated (↑129%, *p* < 0.001) and stimulated (↑25%, *p* = 0.001) saliva of obese teenagers was considerably higher than in their overweight peers.

OSI in non-stimulated and stimulated saliva as well as plasma in obese adolescents was significantly higher than in the controls (↑153%, *p* < 0.001; ↑105%, *p* = 0.001; ↑48%, *p* = 0.01, respectively) (Figure 4).

**Figure 4.** Redox status in overweight and obese adolescents as well as healthy controls. C- control, NWS- non-stimulated whole saliva, OB- obese, OSI- oxidative stress index, OWT- overweight, SWSstimulated whole saliva, TAC- total antioxidant capacity, TOS- total oxidative status. Differences statistically significant at: \* *p* < 0.05, \*\* *p* < 0.005, \*\*\* *p* < 0.0005.

#### *3.5. Oxidation Products*

AGEs in non-stimulated and stimulated saliva as well as plasma of both overweight (↑281%, *p* < 0.001; ↑209%, *p* < 0.001; ↑203%, *p* < 0.001, respectively) and obese (↑347%, *p* < 0.001; ↑423%, *p* < 0.001; ↑244%, *p* < 0.001, respectively) adolescents were significantly higher compared to the saliva of the control group. Only AGEs in stimulated saliva of obese adolescents were considerably higher than in overweight adolescents (↑69%, *p* < 0.001).

MDA in non-stimulated and stimulated saliva as well as plasma in both overweight (↑43%, *p*<0.001; ↑63%, *p* = 0.001; ↑41%, *p* < 0.001, respectively) and obese (↑43%, *p* < 0.001; ↑79%, *p* < 0.001; ↑55%, *p* < 0.001, respectively) teenagers were significantly higher than in saliva of control group adolescents.

The concentration of 4-HNE was notably higher in non-simulated and stimulated saliva as well as plasma of overweight (33% *p* = 0.01; 50% *p* < 0.001; 41% *p* = 0.003) and obese adolescents (84% *p* < 0.001; 95% *p* < 0.001; 104% *p* < 0.001) compared to the control group. The concentration of 4-HNE in non-simulated and stimulated saliva as well as plasma of obese adolescents was considerably higher than in overweight adolescents (37% *p* < 0.001; 50% *p* < 0.001; 43% *p* < 0.001). Similarly, the concentration of 8-OHdG in non-stimulated and stimulated saliva as well as plasma of both overweight (53% *p* < 0.001; 25% *p* = 0.04; 62% *p* < 0.001) and obese adolescents (121% *p* < 0.001; 73% *p* < 0.001; 118% *p* < 0.001) was significantly higher than in saliva and plasma of the controls. The 8-OHdG concentration in non-stimulated and stimulated saliva as well as plasma of obese teenagers was considerably higher than in their overweight peers (43% *p* = 0.001, 38% *p* = 0.007; 34% *p* < 0.001) (Figure 5).

**Figure 5.** Protein, lipid, and DNA oxidation products in overweight and obese adolescents as well as healthy controls. AGE- advanced glycation end products, C- control, MDA- malondialdehyde, NWSnon-stimulated whole saliva, OB- obese, OWT- overweight, SWS- stimulated whole saliva, 4-HNE-4-hydroxynoneal protein adduct, 8-OHdG- 8-hydroxy-D-guanosine. Differences statistically significant at: \*\* *p* < 0.005, \*\*\* *p* < 0.0005.

#### *3.6. ROC Analysis*

The diagnostic utility of salivary redox parameters to differentiate children who are overweight from those who are obese is presented in Table 3. For this purpose, ROC curves were generated, and then the area under the curve (AUC) was calculated. Optimal cut-off values were determined for each parameter that ensured high sensitivity with high specificity. The maximum AUC value, from 0 to 1, is a parameter that determines the discriminatory power of the test.

Particular attention should be paid to SOD, CAT, TOS and OSI in NWS, GSH and AGE in SWS, and CAT in erythrocytes—the AUC of which is close to 1.0, which differentiates overweight adolescents from obese ones (Figure 6).



8-OHdG (pg/mg protein) 0.8625 0.7452–0.9798 <0.0001 <3.440 80 80 0.795 0.6501–0.9399 0.0014 <2.226 75 80 0.8075 0.6761–0.9389 0.0009 >2.690 65 70AGE- advanced glycation end products, CAT- catalase, GSH- reduced glutathione, MDA- malondialdehyde, NWS- non-stimulated whole saliva, OSI- oxidative stress index, Px- salivary peroxidase, SOD- superoxide dismutase, SWS- stimulated whole saliva, TAC- total antioxidant capacity, TOS- total oxidative status, UA- uric acid, 4-HNE- 4-hydroxynoneal protein adduct,8-OHdG-8-hydroxy-D-guanosine.

**Figure 6.** Area under the curve (AUC) of selected redox biomarkers in overweight and obese children. AGE- advanced glycation end products, CAT- catalase, GSH- reduced glutathione, NWS- non-stimulated whole saliva, OSI- oxidative stress index, SOD- superoxide dismutase, SWS- stimulated whole saliva, TOS- total oxidative status, UA- uric acid, 4-HNE- 4-hydroxynoneal protein adduct,.

#### *3.7. Correlations*

We showed a positive correlation between erythrocyte and salivary CAT and TAC in overweight and obese adolescents. We also demonstrated a positive correlation between UA content in plasma and non-stimulated/stimulated saliva of the study group patients (Figure 7B). However, we did not observe a saliva–blood correlation of UA in healthy children and adolescents (Figure 7A).

**Figure 7.** Saliva–blood correlations of the analyzed redox biomarkers in healthy controls (**A**) as well as overweight and obese adolescents (**B**). CAT- catalase, NWS- non-stimulated whole saliva, SWSstimulated whole saliva, TAC- total antioxidant capacity, UA- uric acid.

Correlations between BMI and salivary redox biomarkers and presented in Figures 8 and 9. Interestingly, BMI correlates with most salivary antioxidants and oxidative damage products but only in the study group.

**Figure 8.** Correlations between BMI and salivary antioxidants in healthy children (**A**) as well as overweight and obese adolescents (**B**). BMI- body mass index, CAT- catalase, GSH- reduced glutathione, NWS- non-stimulated whole saliva, Px- salivary peroxidase, SOD- superoxide dismutase, SWSstimulated whole saliva, TAC- total antioxidant capacity, UA- uric acid.

**Figure 9.** Correlations between BMI and salivary oxidative damage in healthy children (**A**) as well as overweight and obese adolescents (**B**). BMI- body mass index, AGE- advanced glycation end products, MDA- malondialdehyde, NWS- non-stimulated whole saliva, OSI- oxidative stress index, SWS- stimulated whole saliva, TOS- total oxidative status, 4-HNE- 4-hydroxynoneal protein adduct, 8-OHdG- 8-hydroxy-D-guanosine.

#### **4. Discussion**

This publication is the first to analyze redox balance in the saliva of overweight and obese adolescents. Generally, we demonstrated disturbances in the activity/concentration of antioxidants as well as oxidative stress in non-stimulated and stimulated saliva of both examined groups compared to their peers with normal body weight, with a higher intensity of oxidative modifications in the saliva of obese adolescents.

Excessive body weight is characterized by chronic (low-grade) inflammation with permanently elevated OS. It has been demonstrated that adipose tissue induces the synthesis of proinflammatory cytokines, such as TNFα, IL-1 and IL-6, which increase the generation of ROS and nitrogen radicals by macrophages and monocytes. ROS production promotes inflammation and expression of molecules as well as growth factors by redox-sensitive transcription factors, mainly NF-κB and the NADPH oxidase pathway [33,34]. The inefficiency of antioxidant systems, observed in the plasma

of obese individuals [16], entails oxidative damage to cellular components and development of obesity-related complications.

It has been shown that obesity results in the dysfunction of salivary glands as well as changes in salivary flow and composition [15–17,35]. As saliva is essential for maintaining appropriate functions of the body, such as swallowing, chewing, carbohydrate digestion, healing of the oral mucosa and tooth enamel remineralization, it is not surprising that excessive body weight increases the risk of gingivitis [36], periodontitis [37], caries [38,39] and inflammatory changes in the oral cavity mucosa [40]. Recently, a significant influence of OS has been increasingly emphasized in explaining the pathogenesis of salivary gland lesions in the course of overweight/obesity in adults [16,17,41]. To the best of our knowledge, there have been no studies evaluating redox balance in the oral cavity of overweight and obese adolescents.

The study by Brown et al. [42] demonstrated that failure of antioxidant systems is related to the duration of obesity. Considering that our study covered adolescents aged 11 to 18 with relatively short overweight/obesity history (4/4.3 years, respectively, data not shown), it is not surprising that salivary and plasma TAC (the sum of both enzymatic and non-enzymatic antioxidants) as well as the content of enzymatic antioxidants were elevated in both non-stimulated and stimulated saliva of overweight and obese adolescents. Therefore, the higher activity of enzymatic antioxidants may be an expression of a highly effective antioxidant barrier that has not been exhausted in oxidative stress conditions. On the other hand, a significant increase in salivary TAC and enzymatic antioxidants as well as, generally, plasma biomarkers (despite the decline in SOD activity, which we explain further) can be considered as a positive adaptative response to the increased ROS generation (↑TOS in plasma as well as non-stimulated and stimulated saliva of both study groups). It was demonstrated that decreased activity of antioxidant enzymes in erythrocytes, accompanying increased plasma TAC, is likely to result from cell damage due to an inflammatory process and leakage of enzymes into the extracellular space [43]. On the other hand, it may result from the use of enzymes in the process of ROS control, or from inactivation of enzymes by free radicals [44].

Interestingly, we found a significant positive correlation between erythrocyte and salivary CAT and TAC in overweight and obese adolescents, which suggests that these salivary parameters could assess the general antioxidant status of overweight and obese adolescents.

Uric acid constitutes 40% of the antioxidant potential of saliva [45]; however, it has been found that at high concentrations it can induce and intensify oxidative damage [46]. Obesity has been demonstrated to increase UA concentration by reducing its renal secretion and accumulating metabolites for UA production [47], which was confirmed by our study. We observed an increase in uric acid concentration both in plasma and stimulated saliva of overweight and obese adolescents as well as in non-stimulated saliva of the latter. The negative correlation between the concentrations of UA and protein in non-stimulated saliva of overweight and obese adolescents, as well as the negative correlation between UA content and Px activity in stimulated saliva, suggest that in both groups UA shifts salivary redox balance towards oxidation and does not oxidize hydroxyl or peroxyl radicals, preventing OS. Moreover, we demonstrated a positive correlation between UA levels in plasma and stimulated saliva of overweight and obese subjects. Considering that plasma UA is a strong predictor of future development of type 2 diabetes [48], measurements of salivary UA may be useful in assessing the risk of this disease.

Although changes in the activity/concentration of antioxidants or ROS concentration may suggest redox imbalance, they are not sufficient to determine the existence and extent of OS. The most reliable determinants of oxidative stress are increased concentrations of oxidative damage products [49,50]. There are numerous markers of oxidative damage to biomolecules; in our study, we assessed 4-HNE protein adducts, MDA, AGE and 8-OHdG. These are only few selected markers, which should be taken into account when interpreting the presented results. The use of other OS indicators could change our observations and conclusions.

Our research showed, however, that the overproduction of free radicals exceeds the capabilities of antioxidant systems of the examined adolescents at the central level as well as in both salivary glands, which was observed as increased concentration of oxidative modification products in plasma and non-stimulated as well as stimulated saliva. It should be noted that intensified oxidative modifications reveal a certain tendency (AGE and MDA in plasma and non-stimulated saliva; MDA in stimulated saliva) or are significantly higher (4-HNE, 8-OHdG in plasma and stimulated and non-stimulated saliva, and AGE in stimulated saliva) in obese adolescents compared to their overweight peers. The obtained results prove that oxidative damage occurred in the salivary glands of overweight adolescents, and was more severe in obese subjects. We noted a significant increase of TOS in non-stimulated and stimulated saliva in obese adolescents compared to the overweight group, but observed no such dependence for TAC and antioxidant enzymes (except CAT in non-stimulated saliva). In our opinion, this is a worrying phenomenon and may be evidence of the beginning of subclinical inefficiency of antioxidant mechanisms, leading to the boost of oxidative damage to the salivary glands of obese adolescents.

We demonstrated that in both non-stimulated and stimulated saliva of overweight and obese patients, GSH levels were significantly decreased and considerably lower in the stimulated saliva of patients with obesity compared to overweight ones. Our results suggest that, GSH is the first to be used up, and perhaps forms the first line of defense against free radicals, which can be easily explained. The main function of GSH is maintaining thiol groups of proteins in a reduced state, which is often necessary for preserving the functional activity of these proteins. It has been shown that the most probable primary object of ROS attack is proteins, and, according to this theory, fatty and nucleic acids are protected by proteins and therefore undergo oxidation at a later stage [51]. A mediator of biomolecule damage in cells is the hydroxyl radical (OH**.** ) [52]. It was shown that the percentage share of primary substrates of the OH**.** reaction is 75% proteins, 21% lipids, and 4% DNA, which is related to the specificity of the mechanism of OH**.** radical production in the Fenton reaction [52].

It is worth mentioning that the stimulation of saliva secretion activates parotid glands, while at rest the main source of saliva is the submandibular glands. Therefore, it was assumed that any disturbances in secretion/composition of non-stimulated saliva reflects disturbances of the submandibular gland function, and disturbed secretion/composition of stimulated saliva indicates disturbed activity of parotid glands [53].

The ability of salivary glands of overweight and obese adolescents to secrete saliva at rest was similar to the results of their peers of normal weight. Protein secretion by submandibular glands of the examined adolescents also did not differ from the control group of normal weight. In relation to the stimulated salivary flow, overweight and obese subjects showed reduced salivary secretion compared to the control group, which is consistent with the results of Modéer et al. [38]. Moreover, the dysfunction of parotid glands intensified with the increase in BMI, as obese adolescents produce significantly less stimulated saliva than their overweight peers, and in 8 obese individuals we recorded the salivary flow value of 0.7 mL/min, which is classified as hyposalivation. Modéer et al. [38] claimed that BMI SDS may be a potential factor to exclude subjects with reduced flow of stimulated saliva. Protein concentration in SWS of overweight and obese adolescents was also reduced compared to the controls. The decrease in stimulated saliva secretion and protein concentration is most probably caused by steatosis of the parotid (but not submandibular) glands observed in obese patients, which decreases the number of secretory units (acini and ducts) [54,55]. It can also be assumed that obesity-related inflammatory milieu (up-regulation of proinflammatory cytokines, ROS), similarly to Sjögren's syndrome [56], activates metalloproteinases, thus disrupting the stromal tissue. This phenomenon may disturb the neurotransmission between the residual neural network and residual secretory units, as well as inhibit the response of follicular cells [57,58].

Our study showed reduced salivary production and enhanced oxidative stress in overweight and obese teenagers. However, in obese children, alterations in salivary gland function are more severe than those who are overweight. This is also reflected in the redox status of saliva (greater disturbances in the antioxidant barrier) as well as higher severity of oxidative damage in children with obesity vs. overweight.

#### **5. Conclusions**

Obese and overweight adolescents present impaired systemic and salivary oxidative status in contrast to their normal weight peers.

Both parotid and submandibular salivary glands lose the ability to maintain redox balance at the level observed in the control group, which was shown by an increased level of oxidized biomolecules. However, redox equilibrium in our study was more disturbed in the saliva and plasma of obese adolescents compared to overweight subjects.

Excess of adipose tissue and deficiency of GSH are the main factors responsible for oxidative damage to the salivary glands.

Dysfunction of parotid glands in relation to salivary secretion deepens with the increase of BMI. Dysfunction of mechanisms responsible for protein synthesis/secretion observed at the overweight stage does not worsen with the increase of body weight in adolescents.

Determinations of salivary CAT and TAC could be used to assess the central antioxidant status of overweight and obese adolescents.

**Author Contributions:** Conceptualization, A.Z., M.M.; Data curation, A.Z. and M.M.;A.K., S.Z., Formal analysis, A.Z. and M.M. A.K., S.Z.; Funding acquisition, A.Z.; Investigation, A.Z. and M.M.; Methodology, A.Z., M.M., Material collection: A.K., K.F., M.S., P.K-S., Supervision, A.Z., K.T-J., A.W.; Validation, A.Z. and M.M.; Visualization, A.Z. and M.M.; Writing—original draft, A.Z. and M.M.; Writing—review and editing, A.Z. and M.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from the Medical University of Bialystok, Poland (grant numbers: SUB/1/DN/20/002/1209; SUB/1/DN/20/002/3330).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Salivary Cytokine Biomarker Concentrations in Relation to Obesity and Periodontitis**

#### **Sanna Syrjäläinen 1, Ulvi Kahraman Gursoy 1,\*, Mervi Gursoy 1, Pirkko Pussinen 2, Milla Pietiäinen 2, Antti Jula 3, Veikko Salomaa 3, Pekka Jousilahti <sup>3</sup> and Eija Könönen 1,4**


Received: 25 November 2019; Accepted: 3 December 2019; Published: 5 December 2019

**Abstract:** Systemic low-grade inflammation is associated with obesity. Our aim was to examine the association between obesity and salivary biomarkers of periodontitis. Salivary interleukin (IL)-1-receptor antagonist (IL-1Ra), IL-6, IL-8, IL-10, and tumor necrosis factor (TNF)-α concentrations were measured from 287 non-diabetic obese (body mass index (BMI) of >35 kg/m2) individuals and 293 normal-weight (BMI of 18.5–25 kg/m2) controls. Periodontal status was defined according to a diagnostic cumulative risk score (CRS) to calculate the risk of having periodontitis (CRS I, low risk; CRS II, medium risk; CRS III, high risk). In the whole population, and especially in smokers, higher IL-8 and lower IL-10 concentrations were detected in the obese group compared to the control group, while in non-smoking participants, the obese and control groups did not differ. IL-1Ra and IL-8 concentrations were higher in those with medium or high risk (CRS II and CRS III, *p* < 0.001) of periodontitis, whereas IL-10 and TNF-α concentrations were lower when compared to those with low risk (CRS I). In multivariate models adjusted for periodontal status, obesity did not associate with any salivary cytokine concentration. In conclusion, salivary cytokine biomarkers are not independently associated with obesity and concentrations are dependent on periodontal status.

**Keywords:** obesity; periodontitis; cytokine; inflammation; saliva

#### **1. Introduction**

Obesity is an increasing health problem in developed countries and is a major risk factor for diabetes, cancer, and cardiovascular diseases [1]. Obesity is linked to both local and systemic inflammation [2–5]. In obese subjects, there is an elevated release of proinflammatory cytokines into serum derived from either adipocytes, stromal vascular fraction cells, or immune cells in adipose tissue [3,5,6]. This elevated cytokine release in obese individuals is a reversible condition, since even mild weight loss can reduce serum cytokine levels [7].

There is substantial evidence showing a positive association between obesity or weight gain and periodontitis [8–11]. The underlying mechanisms explaining this association are not completely elucidated, but one proposed mechanism is that the low-grade systemic inflammation related to obesity could expose obese people to infectious diseases [3,4,12]. Pathogenic bacterial biofilms at the gingival margin trigger the initiation of inflammatory processes in periodontal tissues, including the production of chemokines and proinflammatory cytokines [13,14]. Inflammatory cytokines in the periodontium

are low-molecular weight proteins secreted from both periodontal tissue and immune cells [13]. They are the main regulators of inflammation and tissue destruction in periodontitis [15], and their levels in saliva predict periodontal disease progression and remission [16]. To our knowledge, the association between salivary cytokine concentrations and obesity and periodontitis has not been examined in humans so far.

In the present study, we used a novel diagnostic method, the cumulative risk score (CRS), to detect periodontal disease based on three biomarkers in saliva [17]. CRS combines *Porphyromonas gingivalis*, interleukin (IL)-1β, and matrix metalloproteinase (MMP)-8, and categorizes individuals with low, medium, or high risk of having periodontitis. Our hypothesis is that low-grade inflammation related to obesity affects the cytokine biomarker concentrations in saliva. In this context, our cross-sectional study aimed to examine the relation between obesity and periodontitis-associated salivary cytokine concentrations in smoking and non-smoking individuals.

#### **2. Experimental Section**

#### *2.1. Study Population*

The study consisted of 580 individuals aged 25–74 (mean 55.3) years. They were participants of the Dietary, Lifestyle, and Genetic determinant of Obesity and Metabolic syndrome (DILGOM) study, which was an extension of the population-based National FINRISK2007 health survey to investigate more specifically the effects of diet, lifestyle, and genetic factors on obesity and metabolic syndrome [18,19]. Of the 5024 individuals who participated in the DILGOM study, 287 severely obese (body mass index (BMI) <sup>≥</sup>35 kg/m2) and 297 normal-weight (BMI 18.5–25 kg/m2) controls matched for age and smoking status were included. Exclusion criteria for both cases and controls were diabetes, cardiovascular disease, cancer, or medication for hypercholesterolemia. The protocol of the FINRISK2007 survey included questionnaire data on smoking and other health behaviors, socioeconomic background factors, clinical measurements, and venous blood samples. Participants were categorized by their smoking status as current smokers or non-smokers (smokeless for at least the past 6 months). Height was measured to the nearest 0.1 cm and weight to the nearest 0.1 kg.

#### *2.2. Bacterial and Cytokine Measurements from Salivary Samples*

Paraffin-stimulated whole saliva samples were collected by expectoration into plastic tubes from the DILGOM study participants. All samples were stored frozen at −70 ◦C until laboratory analyses. Before analyses, melted samples were gently mixed, and centrifuged at 10,000 × *g* for 5 min. DNA was isolated from the pellet and used in *P. gingivalis* quantification, while the supernatant was used in cytokine determinations.

Salivary concentrations of IL-1β, IL-1 receptor antagonist (IL-1Ra), IL-6, IL-8, IL-10, tumor necrosis factor alpha (TNF–α), and MMP-8 were analyzed with the flow-cytometric Luminex xMAP technique with commercially available kits by Bio-PlexTM 200 (Bio-Rad Laboratories Incorporation, Santa Rosa, CA, USA).

The amounts of *P. gingivalis* were determined with a quantitative real-time PCR (qPCR) assay as previously described [20] with modifications. Reaction mixtures (total volume 20 μL) contained 2 μL of template DNA, 200 nM primers (Thermo Fisher Scientific, Waltham, MA, USA), and 1 × Universal KAPA SYBR FAST qPCR mastermix (KAPA Biosystems, Wilmington, MA, USA) supplemented with 1 × ROX Low reference dye. qPCR analyses were performed with the Mx3005P Real-Time qPCR System (Stratagene, La Jolla, CA, USA) via the following steps: Initial denaturation at 95 ◦C for 3 min, followed by 40 cycles of 3 s at 95 ◦C and 20 s at 60 ◦C. A dissociation curve was generated from one cycle of 1 min at 95 ◦C, then lowering the temperature gradually to 60 ◦C, 30 s at 60 ◦C, then raising the temperature gradually to 95 ◦C, and 30 s at 95 ◦C. The data were analyzed with the Mx3005P Real-Time qPCR System software and the results were presented as genomic equivalents (GE)/mL saliva.

For the standard curve, the whole *P. gingivalis waaA* gene, encoding 3-deoxy-Dmanno-oct-2-ulosonic acid (Kdo) transferase, was cloned to pJet1.2/blunt vector (Thermo Fisher Scientific, Waltham, MA, USA). The cloned fragment was PCR-amplified in the reaction containing 500 nM primers (Fwd-ATGCGATTCCTTTTCAG and Rew-CTATTTCATGATTCGGTG), 200 μM dNTPs, Phusion DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) 0.04 U/μL, 1 × Phusion High-Fidelity Buffer, and 10 ng of chromosomal DNA of *P. gingivalis* strain W50. The cycling conditions followed the manufacturer's instructions. The purified PCR fragment was ligated into pJet1.2/blunt vector with CloneJET PCR Cloning Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions and ligation mixture was transformed into *Escherichia coli* DH5a competent cells. The correct insert was verified by sequencing. The constructed plasmid, pJet1.2/blunt-Pg, was linearized with FastDigest HindIII restriction enzyme (Thermo Fisher Scientific, Waltham, MA, USA) and used for a tenfold dilution series for the qPCR analysis. The plasmid copy number was determined with "DNA Copy Number and Dilution Calculator" (www.thermofisher.com).

#### *2.3. Periodontal Status Assessment Based on Cumulative Risk Score (CRS)*

Periodontal status of the study population (*n* = 580) was defined according to CRS as described in detail by Gursoy et al. [17]. Briefly, for calculating CRS, the salivary concentrations of *P. gingivalis*, IL-1β, and MMP-8 were independently divided into tertiles and each participant was categorized to one of the three tertiles according to the level of the biomarker in the saliva. The person's cumulative score was calculated by multiplication of three biomarkers' tertile values. Based on these calculations, the study participants were categorized into three periodontal status groups, as follows: CRS I: low risk; CRS II: medium risk; CRS III: High risk [17,21,22].

#### *2.4. Statistical Analyses*

All statistical analyses were conducted using the IBM SPSS Statistic 23.0 software (IBM, Armonk, North Castle, NY, USA). In descriptive statistics, continuous variables were reported as means and standard deviations and the differences between the groups were analyzed by independent *t*-test. Categorical variables were reported as the number of individuals and as percentages. Differences between the groups in categorical variables were analyzed using the chi-square test. Statistical significance was set at a *p*-value of < 0.05.

Results of two saliva samples were not included in analyses due to the low sample quality. Due to skewed distribution, salivary cytokine concentrations were reported as medians and interquartile ranges. Differences of salivary cytokine concentrations between the obese and normal-weight participants and a pairwise comparison between the groups among different CRS categories were conducted using the Mann–Whitney *U* test. A *p*-value significance level of < 0.05 was used. Multinomial logistic regression was used to determine whether salivary cytokines were associated with obesity, before and after adjusting the model for smoking and periodontal status (CRS). For multinomial logistic regression, cytokine concentrations were converted into tertiles, and the lowest tertile was used as the reference group.

#### *2.5. Ethical Issues*

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethical Committee of the Hospital District of Helsinki and Uusimaa. Written informed consent was obtained from each participant.

#### **3. Results**

There was no difference in age, gender, or smoking status between the two study groups, whereas mean BMI was significantly higher in the obese (39.0 kg/m2) than in the normal-weight (22.9 kg/m2) groups (*p* < 0.001) (Table 1). Obese individuals were lower educated (*p* = 0.005) and were more likely to have periodontal disease than their controls (*p* = 0.015).

**Table 1.** Characteristics of the study groups according to their weight status. Significant differences (*p* < 0.05) are presented in bold.


*p*-values: Independent samples *T*-test (age, body mass index (BMI)) and chi-square test (gender, smoking status, educational level, periodontitis).

In the whole population and in smokers, the salivary concentrations of IL-8 were higher (whole population *p* = 0.033; smokers *p* = 0.005) and those of IL-10 were lower (whole population *p* = 0.022; smokers *p* = 0.018) in the obese group than in their controls (Table 2). Other salivary cytokine concentrations did not differ according to weight status. In non-smokers, there was no difference in salivary cytokine concentrations between the groups.

When the study participants were stratified only according to their periodontal status and not by their weight, IL-1Ra and IL-8 concentrations were higher in those with medium or high risk (CRS II and CRS III, *p* < 0.001) of periodontitis, whereas IL-10 and TNF-α concentrations were lower when compared to those with the low risk (CRS I). After weight was taken into account, obese individuals with CRS I had lower IL-10 (*p* = 0.043) and TNF-α (*p* = 0.35) concentrations than their normal-weight controls, while obese individuals with CRS III had higher IL-6 concentrations (*p* = 0.011). Other cytokine concentrations in saliva did not differ between the obese and control groups according to their periodontal status (Table 3).


*J. Clin. Med.* **2019** , *8*, 2152

67

IL-6

IL-8

IL-10

TNF-α All values are given as medians (interquartile letters indicate significant differences between the CRS groups in obese and

superscripts:

Mann–Whitney *U* test.

 3.6 (4.3)

 169 (127)

 3.5 (5.8) 14.4 (23.9)

 4.5 (4.7)

 168 (125)

 5.7 (8.1)

 26.4 (26.4)

 ranges). In each CRS group, cytokine

 0.341

 0.440 **0.043** **0.035**

9.8 (14.5)a

 274 (292)a

1.5 (2.8)a

 2.9 (3.3)

 3.0 (3.0)a 284 (242)a

2.0 (3.0)a 12.4 (13.9)a

concentrations

normal-weight

 individuals as follows: a) A difference in CRS I and b) a difference in CRS II. *P*-values and

 were compared between obese and

0.993 0.682 0.078

0.460

 1.3 (2.2)a

8.9 (15.8)a,b normal-weight

 groups (*p*-values). Superscript

573 (443)a,b

4.8 (6.8)a,b

3.0 (5.5)

566 (464)a,b

1.2 (1.9)a,b 10.3 (10.9)a

**0.011**

0.784

0.698

0.841

The multinomial regression model revealed a significant association between obesity and IL-10 concentrations, but the significance was lost after the model was adjusted for smoking and periodontal status (Table 4).

**Table 4.** Associations of salivary IL-1Ra, IL-6, IL-8, IL-10, and TNF-α tertiles with obesity, before and after adjusting the model for smoking and periodontal status (CRS). Significant associations (*p* < 0.05) are presented in bold.


Odds ratios (95% confidence intervals) and *p*-values: Multinomial logistic regression model.

#### **4. Discussion**

The main finding of the present study was that, despite periodontal status being worse in obese individuals (BMI of <sup>≥</sup>35kg/m2) compared to normal-weight controls (BMI of 18.5–25 kg/m2), the association was not consistently reflected in salivary cytokine concentrations. Instead, the concentrations were merely affected by periodontal status rather than obesity. To our knowledge, this was the first study to investigate potential associations of salivary cytokines with obesity by taking periodontal and smoking states into account.

The relatively large sample size, including 287 severely obese and 293 normal-weight individuals, allowed us to make reliable comparisons of cytokine concentrations in saliva between these groups. However, the cross-sectional study design does not provide any information about the causality between periodontitis, obesity, and cytokines. In addition, the relatively large age range of the study population may have displayed an underestimated effect, since the immune response undergoes remodeling with age. To define the presence of periodontal disease, a novel salivary diagnostic tool, CRS, was used [17]. Its capability has been validated in independent populations twice, showing that the CRS index is more strongly associated with moderate to severe periodontitis than any of the salivary biomarkers alone [21,22]. Finally, as part of the sample collection protocol of this survey study, stimulated saliva samples were collected. The protein concentration was higher in the unstimulated saliva samples than the stimulated saliva samples, however, the unstimulated saliva samples possessed greater inter- and intraindividual variation than the stimulated saliva samples [23].

According to the present study, obese individuals expressed enhanced IL-8 and IL-1Ra but reduced IL-10 concentrations in saliva when compared to normal-weight individuals. This finding was observed especially in smokers. Studies dealing with salivary cytokines in relation to weight are sparse. In a recent small-scale study (*n* = 44), TNF-α concentrations in saliva were shown to be significantly higher in obese than non-obese adults [24]. In obese and non-obese children, on the other hand, salivary TNF-α concentrations did not differ [25]. As both obesity and periodontitis are inflammatory states, it was proposed that obesity-related inflammation could predispose obese subjects to periodontal tissue destruction [12,26]. In the present study, obese individuals displayed mainly CRS II and III, indicating worsened periodontal conditions, more frequently than their normal-weight controls. Nevertheless, there were no consistent differences in the salivary cytokine concentrations between the groups.

In periodontitis, cytokines are mainly released from periodontal tissues after pathogenic bacterial recognition [13,14]. There is also some evidence that adipose tissue-derived cytokines act in a paracrine more than an endocrine manner, and hence do not contribute to cytokine concentrations in the oral cavity [27,28]. It is therefore possible that the local infection of the periodontium has such a strong effect on salivary cytokine concentrations that it overpowers the systemic influence of obesity in statistical analyses. Still, obesity may reinforce the inflammatory response to periodontal pathogens in the periodontium, resulting in an increased susceptibility to periodontitis. It is also possible that the link between obesity and periodontitis is not explained by inflammatory factors.

Obesity and periodontitis share other risk factors, including low socioeconomic status [29]. In the present study, obese individuals were less educated than their controls. This is in line with studies linking educational status to health behaviors as well as eating habits to the prevalence of periodontal disease [29–31]. Obesity-related comorbidities, for example, diabetes mellitus, which is a well-known risk factor for periodontitis [26], could also explain the association between obesity and periodontitis. In our study, diabetic patients were excluded, but the obese individuals may still have insulin resistance, a prediabetic state, which is also a proposed risk factor for periodontitis [32].

In addition to obesity, an unhealthy diet causing weight gain plays a role in the obesity-related inflammatory burden [33–35]. In an experimental rodent model, it was shown that obese rats fed a high-fat and high-carbohydrate diet mimicking the Western diet (known as the "cafeteria diet") caused significantly more advanced alveolar bone loss when compared to non-obese counterparts [34]. In another rodent model, it was observed that a diet enriched with saturated fat was associated with higher inflammatory potential and tissue destruction when compared to a diet high in unsaturated fat in obese mice [34]. Therefore, in future studies regarding obesity and periodontitis in humans, it would be of interest to include the dietary composition of obesity in the analyses.

#### **5. Conclusions**

In conclusion, although obese individuals may be prone to have periodontal disease, obesity does not lead to altered cytokine concentrations in saliva. The associations between obesity and salivary cytokine concentrations may be explained by periodontal status and smoking.

**Author Contributions:** Conceptualization, S.S., U.K.G., and V.S.; data curation, M.G., P.P., A.J., V.S., and P.J.; formal analysis, S.S., U.K.G., M.P., and A.J.; funding acquisition, P.P. and V.S.; methodology, M.G., P.P., M.P., P.J., and E.K.; project administration, U.K.G. and E.K.; resources, V.S. and E.K.; supervision, U.K.G. and E.K.; visualization, S.S.; writing—original draft, S.S. and U.K.G.; writing—review and editing, M.G., P.P., M.P., A.J., V.S., P.J., and E.K.

**Funding:** Funding for FINRISK and DILGOM Studies was provided by the National Public Health Institute/KTL (currently National Institute for Health and Welfare/THL) through budgetary funds from the government and a grant from the Academy of Finland. DILGOM Study also received funding from the Finnish Dental Society Apollonia and from the State Research Funding. V.S. was supported by the Finnish Foundation for Cardiovascular Research. PP received funding from the Finnish Dental Society Apollonia.

**Conflicts of Interest:** V.S. participated in a conference trip sponsored by Novo Nordisk and received a honorarium for participating in an advisory board meeting. He also has an ongoing research collaboration with Bayer Ltd (unrelated to the present study). Other authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*
